As a starting point for the thesis we observed that common IR approaches have typically used either full-text indexing or indexing using concepts and, moreover, that few methods exist where the two are combined in a principled manner. Recent advances in the language modeling for IR framework have enabled the use of rich query representations in the form of query language models. This, in turn, has enabled the use of the natural language associated with concepts to be included in the retrieval model in a principled and transparent manner. We have investigated how we can employ the actual use of concepts as measured by the natural language that people use when they discuss them. Furthermore, recent developments in the semantic web community, such as DBpedia and the inception of the Linked Open Data cloud, have enabled the association of texts with concepts on a large scale. These developments enable us to move beyond manually assigned concepts in domain-specific contexts and into the general domain.
The main motivation for this thesis has been to verify whether knowledge captured in concept languages and the associations between concepts and natural language texts can be successfully used to inform IR algorithms and improve information access. Such algorithms are able to match queries and documents not only on a textual, but also on a semantic level. We present and evaluate several models and methods and perform and report on extensive experiments. In sum, we have shown that employing the (natural) language use associated with concepts can successfully and significantly improve information access.
In the remainder of this chapter we conclude the thesis. We first answer the research questions governing the preceding chapters (Section 8.1) and then conclude the thesis by discussing several directions for future work.